This invention relates to immunoassays for detecting VEGF that can be used as diagnostic and prognostic methods for patients with cancer, cardiovascular, or other pathologies.

It is now well established that angiogenesis is implicated in the pathogenesis of a variety of disorders. These include solid tumors, intra-ocular neovascular syndromes such as proliferative retinopathies or age-related macular degeneration (AMD), rheumatoid arthritis, and psoriasis (Folkman et al. J. Biol. Chem. 267:10931-10934 (1992); Klagsbrun et al. Annu. Rev. Physiol. 53:217-239(1991); and Garner A. Vascular diseases. In: Pathobiology of ocular disease. A dynamic approach. Garner A. Klintworth GK. Eds. 2nd Edition (Marcel Dekker, NY, 1994), pp 1625-1710). In the case of solid tumors, the neovascularization allows the tumor cells to acquire a growth advantage and proliferative autonomy compared to the normal cells. Accordingly, a correlation has been observed between density of microvessels in tumor sections and patient survival in breast cancer as well as in several other tumors (Weidner et al. N Engl J Med 324:1-6 (1991); Horak et al. Lancet 340:1120-1124 (1992); and Macchiarini et al. Lancet 340:145-146 (1992)).

The search for positive regulators of angiogenesis has yielded many candidates, including aFGF, bFGF, TGF-α, TGF-β, HGF, TNF-α, angiogenin, IL-8, etc. (Folkman et al., supra, and Klagsbrun et al., supra). The negative regulators so far identified include thrombospondin (Good et al. Proc. Natl. Acad. Sci. USA. 87:6624-6628 (1990)), the 16-kilodalton N-terminal fragment of prolactin (Clapp et al. Endocrinology, 133:1292-1299 (1993)), angiostatin (O'Reilly et al. Cell 79:315-328 (1994)), and endostatin (O'Reilly et al. Cell 88:277-285 (1996)).

Work done over the last several years has established the key role of vascular endothelial growth factor (VEGF) in the regulation of normal and abnormal angiogenesis (Ferrara et al. Endocr. Rev. 18:4-25 (1997)). The finding that the loss of even a single VEGF allele results in embryonic lethality points to an irreplaceable role played by this factor in the development and differentiation of the vascular system (Ferrara et al., supra).

Furthermore, VEGF has been shown to be a key mediator of ncovascularization associated with tumors and intra-ocular disorders (Ferrara et al., supra). The VEGF mRNA is overexpressed by the majority of human tumors examined (Berkman et al. J Clin Invest 91:153-159 (1993); Brown et al. Human Pathol., 26:86-91 (1995); Brown et al. Cancer Res. 53:4727-4735 (1993); Mattern et al. Brit. J. Cancer. 73:931-934 (1996); and Dvorak et al. Am J. Pathol. 146:1029-1039 (1995)). Also, the concentration of VEGF in eye fluids is highly correlated to the presence of active proliferation of blood vessels in patients with diabetic and other ischemia-related retinopathies (Aiello et al. N. Engl. J. Med. 331:1480-1487 (1994)). Furthermore, studies have demonstrated the localization of VEGF in choroidal neovascular membranes in patients affected by acute macular degeneration (AMD) (Lopez et al. Invest. Ophralmo. Vis. Sci. 37:855-868 (1996)).

VEGF is a heparin binding growth factor with a molecular weight of 45 kD (Plouet et al. EMBO J. 8:3801 (1989); Neufeld et al. Prog. Growth Factor Res. 5:89 (1994)). It is a dimeric glycoprotein consisting of two identical subunits. Although VEGF is encoded from a single gene, at least five isoforms exist in vivo due to alternative mRNA splicing. These isoforms, VEGF121, VEGF145, VEGF165, VEGF189, and VEGF206, contain 121, 145, 165, 189, and 206 amino acids, respectively (Leung et al. Science 246:1306 (1989); Houck et al. Mol. Endocrinol. 5:1806 (1991): Tischer et al. J. Biol. Chem. 266:11947 (1991); Neufeld et al. The FASEB Journal 13:9-22 (1999)). The VEGF isoforms show differing affinities for heparin; VEGF121 binds heparin weakly, while VEGF165, VEGF189, and VEGF206 bind heparin with increasing affinity. VEGF121 and VEGF165 are secreted and both isoforms are found in the circulation. In contrast, VEGF189 and VEGF206 are found mostly associated with heparin sulfate containing proteoglycans in the extracellular matrix (Houck et al. J. Biol. Chem. 267:26031 (1992); Park et al. Mol. Biol. Cell 4:1317 (1993)). Of the five isoforms, VEGF165 is the most abundantly expressed variant in the majority of cells and tissues.

Five receptors for VEGF have been identified: VEGFR-1 (FLT-1), VEGFR-2 (KDR/FLK-1), and VEGFR-3, which are all signaling tyrosine kinases, and Neuropilin-1 and Neuropilin-2, which are both accessory-isoform-specific receptors that bind selectively to VEGF165 (de Vries et al. Science 255:989 (1992); Terman et al. Biochem. Biophys. Res. Commun. 187:1579 (1992); Millauer et al. Cell 72:835 (1993); Neufeld et al., supra). The various roles of these receptors in VEGF biology are under active investigation by numerous groups.

VEGF is produced by tissues and docs not have to enter the circulation to exert its biological effect, but rather acts locally as a paracrine regulator. This raises the question of the significance of circulating VEGF and what role it plays in normal biology or pathology. A recent study by Yang et al. J. Pharm. Exp. Ther. 284:103 (1998) found the clearance of rhVEGF165 from the circulation to be very rapid, suggesting endogenous VEGF in the circulation is most likely the result of continual synthesis of VEGF. In addition, several studies have tried to correlate levels of circulating VEGF with tumor burden and have suggested VEGF levels as a potential prognostic marker (Ferrari and Scagliotti Eur. J. Cancer 32A:2368 (1996); Gasparini et al. J. Natl. Cancer Inst. 89:139 (1997); Kohn Cancer 80:2219 (1997); Baccala et al. Urology 51:327 (1998); Fujisaki et al. Am. J. Gastroenterol. 93:249 (1998)). Clearly the ability to accurately measure VEGF will be important to understand its potential role(s) in many biological processes, such as maintenance of vascular patency, menstrual cycle, ischemia, diabetes, and cancer.

The literature reports widely varying concentrations of endogenous VEGF in normal and diseased patients, ranging from undetectable to high levels. It has been reported that VEGF 165/165 can be proteolytically cleaved into three other forms: a 165/110 heterodimer, a 110/110 homodimer, and a 55-amino-acid C-terminal fragment (Keyt et al. J. Biol. Chem. 271:7788-7795 (1996); Keck et al. Arch. Biochem. Biophys. 344:103-113 (1997)).

The ability to measure endogenous VEGF levels depends on the availability of sensitive and specific assays. Colorimetric, chemiluminescence, and fluorometric based enzyme-linked immunosorbent assays (ELISAs) for VEGF have been reported. Houck et al., supra, (1992); Yeo et al. Clin. Chem. 38:71 (1992); Kondo et al. Biochim. Biophys. Acta 1221:211 (1994); Baker et al. Obstet. Gynecol. 86:815 (1995); Hanatani et al. Biosci. Biotechnol, Biochem. 59:1958 (1995); Leith and Michelson Cell Prolif. 28:415 (1995); Shifren et al. J. Clin. Endocrinol. Metab. 81:3112 (1996); Takano et al. Cancer Res. 56:2185 (1996); Toi et al. Cancer 77:110 (1996); Brekken et al. Cancer Res. 58:1952 (1998); Obermair et al. Br. J. Cancer 77:1870-1874 (1998); Webb et al. Clin. Sci. 94:395-404 (1998). Similar ELISAs have been successfully applied in the determination of low amounts of drugs and other antigenic components in plasma and urine samples, involve no extraction steps, and are simple to carry out.

The Houck et al., supra (1992) describe a colorimetric ELISA that appears to have ng/ml sensitivity, clearly not sensitive enough to detect endogenous VEGF levels. Yeo et al., supra (1992) describe a two-site time-resolved immunofluorometric assay, however, no VEGF was detected in normal sera (Yeo et al. Cancer Res. 53:2912 (1993)). Baker et al., supra (1995), using a modified version of this immunofluorometric assay, reported detectable levels of VEGF in plasma from pregnant women, with higher levels observed in women with preeclampsia. Similar data in pregnant women were reported by Anthony et al. Ann. Clin. Biochem. 34:276 (1997) using a radioimmunoassay. Hanatani et al., supra (1995) developed a chemiluminescent ELISA capable of measuring circulating VEGF and report VEGF levels in sera from 30 normal individuals (male and female) from 8-36 pg/ml. Brekken et al. supra (1998) described ELISA assays using antibodies having binding preference to other the VEGF alone or the VEGF:Flk-1 complex.

An ELISA kit for VEGF detection is commercially available from R&D Systems (Abingdon. U.K.). The R&D VEGF ELISA kit has been used in sandwich assays wherein a monoclonal antibody is used to capture the target VEGF antigen and a polyclonal antibody is used to detect the VEGF. Webb et al. supra (1998). It is not clear whether the detection results using the R&D ELISA kit are influenced by the presence of protcolytical processes. or degradation of VEGF, or by interference of other scrum proteins. Obermair et al., supra (1998).

Keyt et al. J. Biol. Chem. 271:7788-7795 (1996); Keyt et al. J. Biol. Chem. 271:5638 (1996); and Shifren et al., supra (1996) also developed a colorimetric ELISA based on a dual monoclonal antibody pair. Although this ELISA was able to detect elevated VEGF levels in cancer patients, it lacked the sensitivity needed to measure endogenous levels of VEGF in normal individuals. Rodriguez et al. J. Immunol. Methods 219:45 (1998) described a two-site fluorimetric VEGF ELISA that yields a sensitivity of 10 pg/ml VEGF in neat plasma or serum. However, this fluorimetric assay can only detect fully intact 165/165 and 165/110 species of VEGF.

Thus, there is a need to develop a diagnostic and prognostic assay that detects higher measurable levels of VEGF in a biological sample of an animal model or patient than existing ELISAs, and can measure all the isoforms of VEGF.

A multi-site antibody-sandwich ELISA method and kits for VEGF as antigen were developed to detect VEGF in biological samples and used as a diagnostic/prognostic index. Compared to the currently used VEGF ELISAs, the present assay has high sensitivity and is capable of detecting most of the isoforms of endogenous VEGF in circulation.

Specifically, the invention provides a method for detecting VEGF in a biological sample, preferably from vascular, diabetic, or cancer patients, comprising the steps of:

• (a) contacting and incubating the biological sample with pre-mixed capture reagents immobilized to a solid support, wherein the capture reagents are polyclonal and monoclonal antibodies against human VEGF, said monoclonal antibody binding specifically to the C-terminal (residues 111-165) of human VEGF;
• (b) separating the biological sample from the immobilized capture reagents;
• (c) contacting the immobilized capture reagents with a detectable antibody that binds to the KDR and FLT1 receptor binding domains of VEGF; and
• (d) measuring the level of VEGF bound to the capture reagents using a detection means for the detectable antibody.

Preferably, the capture reagents are immobilized in a weight ratio of about 0.8:1 to 1.2:1 of monoclonal to polyclonal antibody. More preferably, the weight ratio is about 1:1 of monoclonal to polyclonal antibody.

In another aspect, the invention provides an immunoassay kit for detecting VEGF in a biological sample, the kit comprising:

• (a) as capture reagents, polyclonal and monoclonal antibodies against human VEGF premixed in a weight ratio of about 0.8:1 to 1.2:1 of monoclonal to polyclonal antibody, wherein the monoclonal antibody binds specifically to the C-terminal (residues 111-165) of human VEGF: and
• (b) as detection reagent, a detectable antibody that binds to the KDR and FLT1 receptor binding domains of VEGF.

The assay herein is unique in that it uses a polyclonal/monoclonal antibody mixture as the capture reagents, and the capture monoclonal antibody binds to the C-terminal portion of VEGF. Most of the previously disclosed VEGF ELISAs are based on either a dual monoclonal antibody pair for capture/detection, or a monoclonal antibody as capture reagent and a polyclonal antibody for detection. If a polyclonal antibody is used alone as the capture antibody, all sensitivity of the assay is lost. The ability of the monoclonal capture antibody to bind the VEGF C-terminus ensures that all the endogenous VEGF molecules, including 165/165, 165/110 and 110/110 can be detected by the assay described herein. Furthermore, the detection antibody of the invention binds to the biologically active regions of VEGF. i.e., the binding domains for the KDR and FLT1 receptors of VEGF, which ensures that the detected VEGF molecules are free from being blocked by, for example, soluble VEGF receptors in the circulation. As such, the assay described herein provides a more accurate measurement of circulating VEGF molecules that are most likely biologically active.

• FIG. 1 shows a comparison of two different preparations of affinity-purified rabbit polyclonal antibody against rhVEGF. with the squares depicting the preferred antibody and the circles representing the same antibody from a different bleed.
• FIG. 2 shows acomparison of ELISAs herein using guanidine versus glycine elution of the affinity-purified rabbit polyclonal antibody against VEGF.
• FIG. 3 shows a comparison of typical standard curves for, and possible hook effect of, three different VEGF ELISAs, wherein the circles show the one-site assay with MAb 3.5F8 alone as coat and detection agent, the squares show the two-site assay with MAb 3.5F8 as coat and MAb A4.6.1 as detection agent, and the diamonds show the multi-site assay herein with MAb 3.5F8 and an affinity-purified polyclonal antibody as coat and MAb A4.6.1 as detection agent.
• FIG. 4 shows monoclonal antibody MAb 3.5F8 coat maximization wherein the ELISA uses 0.4 (circles), 1 (squares), 2 (diamonds), or 4 (triangles) µg/ml monoclonal antibody and 1 µg/ml affinity-purified polyclonal antibody.
• FIG. 5 shows rabbit polyclonal antibody coat maximization wherein the ELISA uses 0 (circles), 0.1 (squares), 0.4 (diamonds), 1 (triangles), 2 (reverse triangles with dotted lines), or 4 (reverse triangles with solid lines) µg/ml affinity-purified polyclonal antibody and 0.4 µg/ml MAb 3.5F8.
• FIG. 6 shows the effect of pH on the multi-site VEGF ELISA herein, wherein the circles represent the ELISA at pH 4, the squares, pH 5, the diamonds, pH 6, the triangles, pH 7, the half-line diamonds, pH 8, and the reverse diamonds, pH 9.
• FIG. 7 shows dilution linearity of six normal human EDTA plasma samples spiked with rhVEGF in the multi-site VEGF ELISA.
• FIGS. 8A-8C show linearity of normal rat EDTA plasma samples spiked with rhVEGF. wherein Fig. 8A shows high spike. Fig. 8B shows mid-spike, and Fig. 8C shows low spike, and wherein the circles are rat pool 1. the squares are rat pool 2. and the diamonds are rat 1.
• FIGS. 9A-9B show linearity of, respectively, four female and four male Yorkshire pig EDTA plasma samples spiked with rhVEGF.
• FIGS. 10A-10C show the specificity of three different ELISA assays for VEGF forms 165/165 (circles), 165/110 (squares). 121/121 (diamonds), and 110/110 (triangles). Fig. 10A shows the specificity of the single-site ELISA using MAb 3.5F8 as coat and detection antibody (Fig. 10A). Fig. 10B shows the specificity of the two-site fluorimetric VEGF ELISA assay using MAb 3.5F8 as coat and MAb A4.6.1 as detection antibody, and Fig. 10C shows the specificity of the multi-site VEGF ELISA herein using MAb 3.5F8 and PAb as coat antibodies and MAb A4.6.1 as detection antibody.
• FIGS. 11A and 11B show, respectively, normal human plasma and serum VEGF detected by a two-site ELISA using only MAb 3.5F8 as coat reagent and by the multi-site ELISA herein using both the PAb and MAb 3.5F8 as coat reagents.
• FIG. 12 shows the amounts of plasma VEGF in cardiac patients using all three assays described in the legend to FIG. 10, where the circles represent the single-site assay, the squares represent the two-site assay, and the triangles represent the multi-site assay.
• FIG. 13 shows plasma VEGF levels in normal donors and cardiovascular patients using the two-site assay with MAb 3.5F8 as coat and MAb A4.6.1 as detection antibody or the multi-site assay herein using MAb 3.5F8 and affinity-purified polyclonal antibody as coat and MAb A4.6.1 as detection antibody, where N is the number of patients.
• FIG. 14 shows serum VEGF from lung cancer patients detected by a two-site ELISA using only MAb 3.5F8 as coat reagent and by the multi-site ELISA herein using both the PAb and MAb 3.5F8 as coat reagents.
• FIG. 15 shows serum VEGF levels in normal donors and diabetic patients (non-insulin-dependent diabetes mellitus (NIDDM) and insulin-dependent diabetes mellitus (IDDM)) using the two-site ELISA with MAb 3.5F8 as coat and MAb A4.6.1 as detection antibody.
• FIGS. 16A and 16B show graphs comparing an affinity-purified polyclonal antibody to DNase and an affinity-purified polyclonal antibody to VEGF as one of the coat reagents in the multi-site assay herein, as compared to the two-site assay using MAb 3.5F8 as coat reagent for human plasma. Fig. 16A shows correlation of plasma VEGF measured by the two-site ELISA using MAb 3.5F8 as the coat reagent (x-axis) with plasma VEGF measured by a multi-site assay with MAb 3.5F8 and a polyclonal antibody to DNase as the coat reagents (filled circles), and with plasma VEGF measured by a multi-site assay as set forth herein using MAb 3.5F8 and the PAb to VEGF as the coat reagents (open circles). Fig. 16B shows the standard curves of the two-site ELISA (filled circles), the multi-site ELISA with MAb 3.5F8 and a polyclonal antibody to DNase as the coat reagents (filled diamonds), and a multi-site assay as set forth herein using MAb 3.5F8 and the PAb to VEGF as the coat reagents (filled squares).

The term "VEGF" as used herein refers to the 165-amino acid vascular endothelial cell growth factor, and related 121-, 145-, 189-, and 206-amino acid vascular endothelial cell growth factors, as described by Leung et al. Science 246:1306 (1989), Houck et al. Mol. Endocrin. 5:1806 (1991), and Neufeld et al., supra, together with the naturally occurring allelic and processed forms of those growth factors.

The term "detecting" is used in the broadest sense to include both qualitative and quantitative measurements of a target molecule. In one aspect, the detecting method as described herein is used to identify the mere presence of VEGF in a biological sample. In another aspect, the method is used to test whether VEGF in a sample is at a detectable level. In yet another aspect, the method can be used to quantify the amount of VEGF in a sample and further to compare the VEGF levels from different samples.

The term "biological sample" refers to a body sample from any animal, but preferably is from a mammal, more preferably from a human. Most preferably, such biological sample is from vascular, diabetic, or cancer patients. Such samples include biological fluids such as serum, plasma, vitreous fluid, lymph fluid, synovial fluid, follicular fluid, seminal fluid, amniotic fluid, milk, whole blood, urine, ccrebro-spinal fluid, saliva, sputum, tcars, perspiration, mucus, and tissue culture medium, as well as tissue extracts such as homogenized tissue, and cellular extracts. The preferred biological sample herein is serum, plasma or urine.

The term "capture reagent" refers to a reagent capable of binding and capturing a target molecule in a sample such that under suitable condition, the capture reagent-target molecule complex can be separated from the rest of the sample. Typically, the capture reagent is immobilized or immobilizable. In a sandwich immunoassay, the capture reagent is preferably an antibody or a mixture of different antibodies against a target antigen.

The term "detectable antibody" refers to an antibody that is capable of being detected either directly through a label amplified by a detection means, or indirectly through, e.g., another antibody that is labeled. For direct labeling, the antibody is typically conjugated to a moiety that is detectable by some means. The preferred detectable antibody is biotinylated antibody.

The term "detection means" refers to a moiety or technique used to detect the presence of the detectable antibody in the ELISA herein and includes detection agents that amplify the immobilized label such as label captured onto a microtiter plate. Preferably, the defection means is a fluorimetric detection agent such as avidin or streptavidin.

The term "antibody" is used in the broadest sense and includes monoclonal antibodies (including agonist, antagonist, and neutralizing antibodies), polyclonal antibodies, multivalent antibodies, multispecific antibodies, and antibody fragments so long as they exhibit the desired binding specificity.

The term "monoclonal antibody" as used herein refers to an antibody obtained from a population of substantially homogeneous antibodies, i.e., the individual antibodies comprising the population are identical except for possible naturally-occurring mutations that may be present in minor amounts. Monoclonal antibodies are highly specific, being directed against a single antigenic site. Furthermore, in contrast to conventional (polyclonal) antibody preparations that typically include different antibodies directed against different determinants (epitopes), each monoclonal antibody is directed against a single determinant on the antigen. The modifier "monoclonal" indicates the character of the antibody as being obtained from a substantially homogeneous population of antibodies, and is not to be construed as requiring production of the antibody by any particular method. For example, the monoclonal antibodies to be used in accordance with the present invention may be made by the hyhridoma method first described by Kohler et al. Nature 256:495 (1975), or may be made by recombinant DNA methods (see, e.g., U.S. Patent No. 4,816,567). The "monoclonal antibodies" may also be isolated from phage antibody libraries using the techniques described in Clackson et al. Nature 352:624-628 (1991) and Marks et al. J. Mol. Biol. 222:581-597 (1991), for example.

The monoclonal antibodies herein specifically include "chimeric" antibodies (immunoglobulins) in which a portion of the heavy and/or light chain is identical with or homologous to corresponding sequences in antibodies derived from a particular species or belonging to a particular antibody class or subclass, while the remainder of the chain(s) is identical with or homologous to corresponding sequences in antibodies derived from another species or belonging to another antibody class or subclass, as well as fragments of such antibodies, so long as they exhibit the desired biological activity (U.S. Patent No. 4,816,567; and Morrison et al. Proc. Natl. Acad. Sci. USA 81:6851-6855 (1984)).

"Humanized" forms of non-human (e.g., murine) antibodies are chimeric antibodies that contain minimal sequence derived from non-human immunoglobulin. For the most part, humanized antibodies are human immunoglobulins (recipient antibody) in which hypervariable region residues of the recipient are replaced by hypervariable region residues from a non-human species (donor antibody) such as mouse, rat, rabbit or nonhuman primate having the desired specificity, affinity, and capacity. In some instances, framework region (FR) residues of the human immunoglobulin are replaced by corresponding non-human residues. Furthermore, humanized antibodies may comprise residues that are not found in the recipient antibody or in the donor antibody. These modifications are made to further refine antibody performance. In general, the humanized antibody will comprise substantially all of at least one, and typically two, variable domains, in which all or substantially all of the hypervariable regions correspond to those of a non-human immunoglobulin and all or substantially all of the FRs are those of a human immunoglobulin sequence. The humanized antibody optionally also will comprise at least a portion of an immunoglobulin constant region (Fc), typically that of a human immunoglobulin. For further details, see Jones et al. Nature 321:522-525 (1986); Rcichmann et al. Nature 332:323-329 (1988); and Presta Curr. Op. Struct. Biol. 2:593-596 (1992).

"Mammal" for purposes of treatment refers to any animal classified as a mammal, including humans, domestic, and farm animals, and zoo, sports, or pet animals, such as dogs, horses, cats, sheep, pigs, cows, etc. Preferably, the mammal is human.

The terms "cancer", "cancerous", and "malignant" refer to or describe the physiological condition in mammals that is typically characterized by unregulated cell growth. Examples of cancer include but are not limited to, carcinoma including adenocarcinoma, lymphoma, blastoma, melanoma, sarcoma, and leukemia. More particular examples of such cancers include squamous cell cancer, small-cell lung cancer, non-small cell lung cancer, gastrointestinal cancer. Hodgkin's and non-Hodgkin's lymphoma, pancreatic cancer, glioblastoma, cervical cancer, ovarian cancer, liver cancer such as hepatic carcinoma and hepatoma, bladder cancer, breast cancer, colon cancer, coloreetal cancer, endometrial carcinoma, salivary gland carcinoma, kidney cancer such as renal cell carcinoma and Wilms' tumors, basal cell careinoma, melanoma, prostate cancer, vulval cancer, thyroid cancer, testicular cancer, esophageal cancer, and various types of head and neck cancer. The preferred cancers for treatment herein are breast, colon, lung, and melanoma.

The phrases "vascular" and "cardiovascular" are used interchangeably and describe patients with indications that stimulate angiogenesis and/or cardiovascularization, and those that inhibit angiogenesis and/or cardiovascularization. Such disorders include, for example, arterial disease, such as atherosclerosis, hypertension, inflammatory vasculitis, Reynaud's disease and Reynaud's phenomenon, ancurysms, and aricrial restenosis: venous and lymphatic disorders such as thrombophlebitis, lymphangitis, and lymphedema; and other vascular disorders such as peripheral vascular disease, cancer such as vascular tumors, e.g., hemangioma (capiliary and cavernous), glomus tumors, telangiectasia, bacillary angiomatosis, hemangioendothelioma, angiosarcoma, haemangiopericytoma, Kaposi's sarcoma, lymphangioma, and lymphangiosarcoma, tumor angiogenesis, trauma such as wounds, bums, and other injured tissue, implant fixation, scarring, ischemia reperfusion injury, rheumatoid arthritis, cerebrovascular disease, renal diseases such as acute renal failure, and osteoporosis. This would also include angina, myocardial infarctions such as acute myocardial infarctions, cardiac hypertrophy, and heart failure such as congestive heart failure (CHF).

The term "diabetes" refers to a progressive disease of carbohydrate metabolism involving inadequate production or utilization of insulin and is characterized by hyperglycemia and glycosuria. This term includes all forms of diabetes, such as type I and type II diabetes and insulin-resistant diabetes, such as Mendenhall's Syndrome. Werner Syndrome, leprechaunism, lipoatrophic diabetes, and other lipoatrophies.

The term "affinity purified" refers to purifying a substance by eluting it through an affinity chromatography column.

The assay described herein is a multi-site immunoassay utilizing the following steps.

In the first step of the assay herein, the biological sample is contacted and incubated with the immobilized capture (or coat) reagents, which are an anti-VEGF monoclonal antibody and a polyclonal antibody directed against VEGF. These antibodies may be from any species, but preferably the monoclonal antibody is a murine or rat monoclonal antibody, more preferably murine, and most preferably MAb 3.5F8 (Rodrigeuz et al., supra (1998)), and the polyclonal antibody is an anti-rabbit or anti-goat antibody, more preferably anti-rabbit. Furthermore, the polyclonal antibody is preferably affinity purified, to decrease background. Hence, in a specific preferred embodiment, the immobilized monoclonal antibody is a murine monoclonal antibody, most preferably MAb 3.5F8, and the immobilized polyclonal antibody is an affinity-purified rabbit antibody. The immobilized capture reagents are mixed together before they are immobilized. Immobilization conventionally is accomplished by insolubilizing the capture reagents either before the assay procedure, as by adsorption to a water-insoluble matrix or surface (U.S. Pat. No. 3.720.760) or non-covalent or covalent coupling (for example, using glutaraldehyde orcarbodiimide cross-linking, with or without prior activation of the support with, e.g., nitric acid and a reducing agent as described in U.S. Pat. No. 3.645.852 or in Rotmans et al. J. Immunol. Methods 57:87-98 (1983)), or afterward. e.g., by immunoprecipitation.

The solid phase used for immobilization may be any inert support or carrier that is essentially water insoluble and useful in immunomctric assays, including supports in the form of, e.g., surfaces, particles, porous matrices, etc. Examples of commonly used supports include small sheets, Sephadex, polyvinyl chloride, plastic beads, and assay plates or test tubes manufactured from polyethylene, polypropylene, polystyrene, and the like including 96-well microtiter plates, as well as particulate materials such as filter paper, agarose, cross-linked dextran, and other polysaccharides. Alternatively, reactive water-insoluble matrices such as cyanogen bromide-activated carbohydrates and the reactive substrates described in U.S. Pat. Nos. 3,969,287; 3,691,016; 4,195,128; 4,247,642; 4,229,537; and 4,330,440 are suitably employed for capture reagent immobilization. In a preferred embodiment the immobilized capture reagents are coated on a microtiter plate, and in particular the preferred solid phase used is a multi-well microtiter plate that can be used to analyze several samples at one time. The most preferred is a microtest 96-well ELISA plate such as that sold as Nune Maxisorb or Immulon.

The solid phase is coated with the pre-mixcd capture reagents as defined above, which may be linked by a non-covalent or covalent interaction or physical linkage as desired. Techniques for attachment include those described in U.S. Pat. No. 4,376,110 and the references cited therein. If covalent, the plate or other solid phase is incubated with a cross-linking agent together with the capture reagent under conditions well known in the art such as for 1 hour at room temperature.

Commonly used cross-linking agents for attaching the pre-mixed capture reagents to the solid phase substrate include, e.g., 1,1-bis(diazoacetyl)-2-phenylethane, glutaraldehyde, N-hydroxysuccinimide esters, for example, esters with 4-azidosalicylic acid, homobifunctional imidoesters, including disuccinimidyl esters such as 3,3'-dithiobis(succinimidylpropionate), and bifunctional maleimides such as bis-N-maleimido-1,8-octane. Derivatizing agents such as methyl-3-[(p-azidophenyl)dithio]propioimidate yield photoactivatable intermediates capable of forming cross-links in the presence of light.

If 96-well plates are utilized, they are preferably coated with the mixture of capture reagents (typically diluted in a buffer such as 0.05 M sodium carbonate by incubation for at least about 10 hours, more preferably at least overnight, at temperatures of about 4-20°C. more preferably about 4-8°C, and at a pH of about 8-12. more preferably about 9-10. and most preferably about 9.6. If shorter coating times (1-2 hours) are desired, one can use 96-well plates with nitrocellulose filter bottoms (Millipore MULTISCREEN™) or coat at 37°C. The plates may he stacked and coated long in advance of the assay itself, and then the assay can be carried out simultaneously on several samples in a manual, semi-automatic, or automatic fashion, such as by using robotics.

The coated plates are then typically treated with a blocking agent that binds non-speeifically to and saturates the binding sites to prevent unwanted binding of the free ligand to the excess sites on the wells of the plate. Examples of appropriate blocking agents for this purpose include, e.g., gelatin, bovine serum albumin. egg albumin, casein, and non-fat milk. The blocking treatment typically takes place under conditions of ambient temperatures for about 1-4 hours, preferably about 1.5 to 3 hours.

After coating and blocking, the VEGF standard (purified VEGF) or the biological sample to be analyzed, appropriately diluted, is added to the immobilized phase. The preferred dilution rate is about 5-15%, preferably about 10%, by volume. Buffers that may be used for dilution for this purpose include (a) PBS containing 0.5% BSA. 0.05% TWEEN 20™ detergent (P20), 0.05% PROCLIN™ 300 antibiotic, 5 mM EDTA, 0.25% Chaps surfactant, 0.2% beta-gamma globulin, and 0.35M NaCl; (b) PBS containing 0.5% BSA, 0.05% P20, and 0.05% PROCLIN™ 300. pH 7; (c) PBS containing 0.5% BSA, 0.05% P20, 0.05% PROCLIN™ 300, 5 mM EDTA, and 0.35 M NaCl, pH 6.35; (d) PBS containing 0.5% BSA, 0.05% P20, 0.05% PROCLIN™ 300, 5 mM EDTA, 0.2% beta-gamma globulin, and 0.35 M NaCl; and (e) PBS containing 0.5% BSA. 0.05% P20, 0.05% PROCLIN™ 300. 5 mM EDTA. 0.25% Chaps, and 0.35 M NaCl. Buffer (c) is the preferred buffer for the assay herein since it has the best differentiation between each standard as well as the biggest signal-to-noise ratio. PROCLIN™ 300 acts as a preservative, and TWEEN 20™ acts as a detergent to eliminate non-specific binding. The added EDTA and salt of buffer (c) act to decrease the background over the other buffers, including buffer (b).

The weight ratio of the capture reagents (monoclonal antibody to polyclonal antibody) is preferably about 0.8:1 to about 1.2:1, more preferably about 1:1. The amount of capture reagents employed is sufficiently large to give a good signal in comparison with the VEGF standards, but not in molar excess compared to the maximum expected endogenous VEGF level in the sample. For sufficient sensitivity, it is preferred that the amount of biological sample added be such that the immobilized capture reagents are in molar excess of the maximum molar concentration of free VEGF anticipated in the biological sample after appropriate dilution of the sample. This anticipated level depends mainly on any known correlation between the concentration levels of the free VEGF in the particular biological sample being analyzed with the clinical condition of the patient. Thus, for example, cancer patients may have a maximum expected concentration of free VEGF in their serum that is quite high, whereas a normal child or adult will be expected to have a much lower level of free VEGF in their serum based on what is known in the literature.

If too much of the capture reagents are present, however, the capture reagents will compete with the anti-VEGF present in the biological sample for the bound VEGF, yielding inaccurate results. Thus, while the concentration of the capture reagents will generally he determined by the concentration range of interest of the VEGF taking any necessary dilution of the biological sample into account, the final concentration of the capture reagents will normally be determined empirically to maximize the sensitivity of the assay over the range of interest. However, as a general guideline, the molar excess is suitably less than about ten-fold of the maximum expected molar concentration of free VEGF in the biological sample after any appropriate dilution of the sample. Most preferably, the amount of monoclonal antibodies immobilized is about 0.4 µg/ml and the amount of polyclonal antibodies immobilized is about 0.4 µg/ml.

The conditions for incubation of sample and immobilized capture reagent are selected to maximize sensitivity of the assay and to minimize dissociation. Preferably, the incubation is accomplished at fairly constant temperatures, ranging from about 0°C to about 40°C, preferably from about 36 to 38°C to obtain a less variable, lower coefficient of variant (CV) than at, e.g. room temperature. The time for incubation depends primarily on the temperature, being generally no greater than about 10 hours to avoid an insensitive assay. Preferably, the incubation time is from about 0.5 to 3 hours, and more preferably 1.5-3 hours at 36-38°C to maximize binding of free VEGF to capture reagents. The duration of incubation may be longer if a protease inhibitor is added to prevent proteases in the biological fluid from degrading the VEGF.

At this stage, the pH of the incubation mixture will ordinarily be in the range of about 6-9.5, preferably in the range of about 6-7, more preferably about 6.0 to 6.5, and most preferably the pH of the assay (ELISA) diluent is 6.35 ± 0.1. Acidic pH such as pH 4-5 decreased recovery of VEGF. The pH of the incubation buffer is chosen to maintain a significant level of specific binding of the capture reagents to the VEGF being captured. Various buffers may be employed to achieve and maintain the desired pH during this step, including borate, phosphate, carbonate. Tris-HCl or Tris-phosphate, acetate, barbital, and the like. The particular buffer employed is not critical to the invention, but in individual assays one buffer may be preferred over another.

In the second step of the assay method herein, the biological sample is separated (preferably by washing) from the immobilized capture reagents to remove uncaptured VEGF. The solution used for washing is generally a buffer ("washing buffer") with a pH determined using the considerations and buffers described above for the incubation step, with a preferable pH range of about 6-9. The washing may be done three or more times. The temperature of washing is generally from refrigerator to moderate temperatures, with a constant temperature maintained during the assay period, typically from about 0-40°C. more preferably about 4-30°C. For example, the wash buffer can be placed in ice at 4 °C in a reservoir before the washing, and a plate washer can be utilized for this step. A cross-linking agent or other suitable agent may also be added at this stage to allow the now-bound VEGF to be covalently attached to the capture reagents if there is any concern that the captured VEGF may dissociate to some extent in the subsequent steps.

In the next step, the immobilized capture reagents are contacted with detectable antibodies, preferably at a temperature of about 20-40°C, more preferably about 36-38°C, with the exact temperature and time for contacting the two being dependent primarily on the detection means employed. For example, when 4-methylumbelliferyl-β-galactoside (MUG) and streptavidin-β-galactosidase are used as the means for detection, preferably the contacting is carried out overnight (e.g., about 15-17 hours or more) to amplify the signal to the maximum. While the detectable antibody may he a polyclonal or monoclonal antibody, preferably it is a monoclonal antibody, more preferably murinc, and most preferably MAb A4.6.1. Also, the preferred detectable antibody is directly detectable, and preferably has a fluorimetric label. The fluorimetric label has greater sensitivity to the assay compared to the conventional colorimetric label. More preferably, the delectable antibody is biotinylated and the detection means is avidin or streptavidin-β-galactosidase and MUG.

Preferably a molar excess of an antibody with respect to the maximum concentration of free VEGF expected (as described above) is added to the plate after it is washed. This antibody (which is directly or indirectly detectable) is preferably a polyclonal antibody, although any antibody can he employed. The affinity of the antibody must be sufficiently high that small amounts of the free VEGF can be detected, but not so high that it causes the VEGF to be pulled from the capture reagents.

In the last step of the assay method, the level of free VEGF that is now hound to the capture reagents is measured using a detection means for the detectable antibody. If the biological sample is from a vascular, diabetic, or cancer patient, the measuring step preferably comprises comparing the reaction that occurs as a result of the above three steps with a standard curve to determine the level of VEGF compared to a normal individual.

Polyclonal antibodies to the VEGF generally are raised in animals by multiple subcutaneous (sc) or intraperitoneal (ip) injections of the VEGF and an adjuvant. It may be useful to conjugate the VEGF or a fragment containing the target amino acid sequence to a protein that is immunogenic in the species to be immunized, e.g., keyhole limpet hemocyanin, serum albumin, bovine thyroglobulin, or soybean trypsin inhibitor using a bifunctional or derivatizing agent, for example, maleimidobenzoyl sulfosuccinimide ester (conjugation through cysteine residues), N-hydroxysuccinimide (through lysine residues), glutaraldehyde, succinic anhydride, SOCl₂, or R¹N = C = NR. where R and R¹ are different alkyl groups.

The antibodies used as the coat or detectable antibodies may be obtained from any convenient vertebrate source, such as murine, primate, lagomorpha, goat, rabbit, rat, chicken, bovine, ovine, equine, canine, feline, or porcine. Chimeric or humanized antibodies may also be employed, as described, e.g., in U.S. Pat. No. 4.816.567; Morrison et al. Proc. Natl. Acad. Sci. USA 81:6851 (1984); Neuberger et al. Nature 312: 604 (1984); Takeda et al. Nature 314:452 (1985); and WO 98/45331 published October 15, 1998, as well as in those additional references set forth above.

Animals may be immunized against the immunogenic conjugates or derivatives by combining 1 mg or 1 µg of conjugate (for rabbits or mice, respectively) with 3 volumes of Freund's complete adjuvant and injecting the solution intradermally at multiple sites. One month later the animals are boosted with 1/5 to 1/10 the original amount of conjugate in Freund's incomplete adjuvant by subcutaneous injection at multiple sites. 7 to 14 days later animals are bled and the serum is assayed for anti-VEGF titer. Animals are boosted until the titer plateaus. Preferably, the animal is boosted with the conjugate of VEGF. but conjugated to a different protein and/or through a different cross-linking agent. Conjugates also can be made in recombinant cell culture as protein fusions. Also, aggregating agents such as alum are used to enhance the immune response. Methods for the production of polyclonal antibodies are described in numerous immunology textbooks, such as Davis et al. Microbiology, 3rd Edition. (Harper & Row. New York, New York. 1980).

Monoclonal antibodies are prepared by recovering spleen cells from immunized animals and immortalizing the cells in conventional fashion, e.g. by fusion with myeloma cells or by Epstein-Barr virus transformation, and screening for clones expressing the desired antibody. See, e.g., Kohler and Milstein Eur. J. Immunol. 6:511 (1976). Monoclonal antibodies, or the antigen-binding region of a monoclonal antibody, such as Fab or (Fab), fragments, may alternatively be produced by recombinant methods.

Examples of suitable antibodies include those already utilized in known RIAs for the protein in question, e.g., those antibodies directed against VEGF as described in the references given in the introduction herein.

The anti body added to the immobilized capture reagents will be either directly labeled, or detected indirectly by addition, after washing off of excess first antibody, of a molar excess of a second, labeled antibody directed against IgG of the animal species of the first antibody. In the latter, indirect assay, labeled antiscra against the first antibody are added to the sample so as to produce the labeled antibody in situ.

The label used for either the first or second antibody is any detectable functionality that docs not interfere with the binding of free VEGF to the antibody. Examples of suitable labels are those numerous labels known for use in immunoassay, including moieties that may be detected directly, such as fluorochrome, chemiluminseent, and radioactive labels, as well as moieties, such as enzymes, that must be reacted or derivatized to be detected. Examples of such labels include the radioisotopes ³²P, ¹⁴C, ¹²⁵I, ³H, and ¹³¹I, fluorophores such as rare earth chelates or fluorescein and its derivatives, rhodamine and its derivatives, dansyl, umbelliferone, luceriferascs, e.g., firefly luciferase and bacterial luciferase (U.S. Pat. No. 4,737.456), luciferin. 2,3-dihydrophthalazinediones, horseradish peroxidase (HRP), alkaline phosphalase, β-galactosidase, glucoamylase, lysozyme, saccharide, oxidases, e.g., glucose oxidase, galactose oxidase, and glucose-6-phosphate dehydrogenase, heterocyclic oxidases such as uricase and xanthine oxidase, coupled with an enzyme that employs hydrogen peroxide to oxidize a dye precursor such as HRP. lactoperoxidase, or microperoxidase, biotin/avidin, biotin/streptavidin, biotin/Streptavidin-β-galactosidase with MUG, spin labels, bacteriophage labels, stable free radicals, and the like. As noted above, the fluorimetric detection is preferred.

Conventional methods are available to bind these labels covalently to proteins or polypeptides. For instance, coupling agents such as dialdehydes, carbodiimides, dimaleimides, bis-imidates, bis-diazotized benzidine, and the like may be used to tag the antibodies with the above-described fluorescent, chemiluminescent, and enzyme labels. See, for example, U.S. Pat. Nos. 3,940,475 (fluorimetry) and 3,645,090 (enzymes); Hunter et al. Nature 144:945 (1962); David et al. Biochemistry 13:1014-1021 (1974); Pain et al. J. Immunol. Methods 40:219-230 (1981); and Nygren J. Histochem. and Cytochem. 30:407-412 (1982). Preferred labels herein are fluorescent to increase amplification and sensitivity to 8 pg/ml, more preferably biotin with streptavidin-β-galactosidase and MUG for amplifying the signal.

The conjugation of such label, including the enzymes, to the antibody is a standard manipulative procedure for one of ordinary skill in immunoassay techniques. See, for example, O'Sullivan et al. "Methods for the Preparation of Enzyme-antibody Conjugates for Use in Enzyme Immunoassay," in Methods in Enzymology, ed. J.J. Langone and H. Van Vunakis, Vol. 73 (Academic Press, New York. New York, 1981), pp. 147-166.

Following the addition of last labeled antibody, the amount of bound antibody is determined by removing excess unbound labeled antibody through washing and then measuring the amount of the attached label using a detection method appropriate to the label, and correlating the measured amount with the amount of free VEGF in the biological sample. For example, in the case of enzymes, the amount of color developed and measured will be a direct measurement of the amount of VEGF present. Specifically, if HRP is the label, the color is detected using the substrate OPD at 490 nm absorbance.

In one example, after an enzyme-labeled second antibody directed against the first unlabeled antibody is washed from the immobilized phase, color or chemiluminiscence is developed and measured by incubating the immobilized capture reagent with a substrate of the enzyme. Then the amount of free VEGF concentration is calculated by comparing with the color or chemituminescence generated by the standard VEGF run in parallel.

As a matter of convenience, the assay method of this invention can be provided in the form of a kit. Such a kit is a packaged combination including the basic elements of:

• (a) capture reagents comprised of polyclonal and monoclonal antibodies against human VEGF molecule, wherein the monoclonal antibody binds specifically to the C-terminal of the VEGF molecule, in a weight ratio of about 0.8:1 to 1.2:1 of monoclonal to polyclonal antibody; and
• (b) detection reagents comprised of detectable (labeled or unlabeled) antibodies that bind to the KDR and FLT1 receptor binding domains of VEGF.

These basic elements are defined hereinabove.

Preferably, the kit further comprises a solid support for the capture reagents, which may be provided as a separate element or on which the capture reagents are already immobilized. Hence, the capture antibodies in the kit may be immobilized on a solid support, or they may be immobilized on such support that is included with the kit or provided separately from the kit. Preferably, the capture reagents are coated on a microtiter plate. The detection reagent may be labeled antibodies detected directly or unlabeled antibodies that are detected by labeled antibodies directed against the unlabeled antibodies raised in a different species. Where the label is an enzyme, the kit will ordinarily include substrates and cofactors required by the enzyme, and where the label is a fluorophore, a dye precursor that provides the detectable chromophore. Where the detection reagent is unlabeled, the kit may further comprise a detection means for the detectable antibodies, such as the labeled antibodies directed to the unlabeled antibodies, preferably in a fluorimetric-detected format.

In a preferred specific embodiment, the weight ratio of monoclonal antibody to polyclonal antibody in the kit is about 1:1, the detectable antibody is a biotinylated murine monoclonal antibody, the monoclonal antibody is murine or rat, more preferably murine, and most preferably MAb 3.5F8, the polyclonal antibody is affinity purified, and more preferably from goat or rabbit, most preferably rabbit, and the amount of murine monoclonal antibodies is 0.4 µg/ml and the amount of rabbit polyclonal antibodies is 0.4 µg/ml. Preferably, the capture reagents are immobilized in this kit. Also, preferably the detectable antibody is MAb A4.6.1.

The kit also typically contains instructions for carrying out the assay, and/or VEGF as an antigen standard (e.g., purified VEGF, preferably recombinantly produced VEGF), as well as other additives such as stabilizers, washing and incubation buffers, and the like.

Examples of standards for VEGF are recombinant human VEGF produced in mammalian cells available from Genentech. Inc., South San Francisco. California.

The components of the kit will be provided in predetermined ratios, with the relative amounts of the various reagents suitably varied to provide for concentrations in solution of the reagents that substantially maximize the sensitivity of the assay. Particularly, the reagents may be provided as dry powders, usually lyophilized, including excipients, which on dissolution will provide for a reagent solution having the appropriate concentration for combining with the sample to be tested.

The following examples are intended to illustrate one embodiment now known for practicing the invention, but the invention is not to be considered limited to these examples.

Purified recombinant human VEGF165 (rhVEGF) expressed in Eschcrichia coli (Genentech, South San Francisco, CA) was used as the standard and for the controls (prepared in ELISA diluent as defined below and stored at -70°C). Streptavidin-β-gafactosidase (Strep-β-gal) was purchased from Boehringer Mannheim. W. Germany; MUG was purchased from Sigma, St. Louis, MO. Dimethylsulfoxide (DMSO) was purchased from Sigma.

Antibodies against rhVEGF165 were prepared as described in Kim et al., Growth Factors, 7:53 (1992). Briefly, BALB/c mice were hyperimmunized intraperitoneally with a 10 mg dose of rhVEGF165 conjugated to keyhole limpet hemocyanin. Spleen cells were fused with a mouse myeloma line and culture supernatants from wells containing hybridomas were screened for the presence of MAbs to rhVEGF165 by an ELISA. Positive hybridomas were cloned twice using the limiting dilution technique. The monoclonal antibodies used in this ELISA have been characterized in Kim et al., supra (1992). One of the capture antibodies, MAb 3.5F8, is thought to bind near the heparin binding domain, amino acid residues 111-165, with a K_d of 13 pM. Rodriguez et al., supra (1998).

The rabbit polyclonal antibody (PAb) used as the other coat antibody was generated by injecting VEGF into a rabbit using a standard protocol, and purified by passing it through an affinity column to which VEGF was coupled to capture the polyclonal antibody, thus removing the immunoglobulins from the sample. The molecules that are not the desired antibody were washed off and the bound antibody was eluted with 0.2 M glycine, pH 2, then the pH was brought to neutral prior to dialysis overnight in PBS at 4°C, and the elutent containing the antibody was used for the multi-site assay.

The detection antibody, MAb A4.6.1, binds rhVEGF165 with a Kd of 86 pM. Several lines of evidence suggest that this MAb binds rhVEGF near the KDR receptor binding region (Kim et al., supra (1992)).

The MAb A4.6.1 was biotinylated with biotinylamidocaproic acid-N-hydroxysuccinimide ester (Biotin-X-NHS) (Research Organics. Cleveland, OH) according to the following protocol. The MAb A4.6.1 was dialyzed against 100 mM NaHCO₃, pH 8.5 overnight at 2-8°C. A total of 60 µl of a 5-mg/ml solution of Biolin-X-NHS in DMSO was added to the MAb (adjusted after dialysis to a concentration of 2-10 mg/ml) using a 1:10 (w/w) ratio of Biotin:MAb. This mixture was allowed to incubate for two hours at ambient temperature with gentle agitation, and the reaction was stopped by the addition of 5 µl of ethanolamine. After conjugation, the antibody was extensively dialyzed against PBS at 2-8°C with gentle agitation and PBS changed every two hours for a total of three times.

Two MAbs. 3.5F8 (coat) and biotinylated A4.6.1 (detection), and one PAb (coat) as described above were used to develop a specific and sensitive VEGF ELISA. In this ELISA, 100 µl/well each of MAh 3.5F8 and the affinity-purified PAb were mixed together and then added to MaxiSorp™ 96-well microtiter plates (Nunc. Roskilde, Denmark) at 0.4 µg/ml each in 0.05 M sodium carbonate, pH 9.6. Following 24-74-hour incubation at 2-8°C, the coated plates were washed 3 times with 400µl ELISA wash buffer (PBS containing 0.05% TWEEN-20™ detergent) using a BIOTEK EL304™ platewasher (Biotek Instruments, Winooski, VT), and blocked with ELISA blocking diluent at 200 µl/well (PBS containing 0.5% BSA, 0.05% TWEEN-20™, and 0.05% PROCLIN™ 300 antibiotic, pH 7.2) for 1-3 hours at ambient temperature with agitation. After blocking, the plates were washed again 3 times with 400µl ELISA wash buffer. Then, 100 µl/well of standards, samples, or controls were added to duplicate wells and incubated for 1.5-2 hours at 37°C with agitation. For quantitation of rhVEGF165 in human plasma, the standard curve was prepared in ELISA diluent (PBS containing 0.5% BSA, 0.05% TWEEN-20™, 0.05% PROCLIN™ 300. 5 mM EDTA. and 0.35 M NaCl, pH 6.3 ± 0.1). The standard curve was 128 pg/ml diluted serially 1:2 to 2 pg/ml. After the sample/standard incubation, the plates were washed six times with 400 µl ELISA wash buffer, and 100 µl/well of MAb A4.6.1-Biotin, freshly diluted 1:200 to its optimal concentration in ELISA diluent, was added to the plates. After a 1.5-2-hour incubation at 37°C with agitation, the plates were washed six times as described above and 100 µl/well of strep-β-gal, diluted 1:40K in ELISA diluent, was added to the plates. After a 45-60-minute incubation at 37°C with agitation, the plates were washed 6 times as described above and 100 µl/well of MUG/DMSO (1/100), freshly diluted to 340 µg/ml in a solution of 0. 1 M NaPO₄, containing 1 mM MgCl₂ at pH 7.3±0.1, was added to the plates. The substrate reaction incubated for 15-17 hours at 37°C with agitation in the dark (plate was wrapped with foil). The reaction was stopped by adding 150 µl/well of 0.15 M glycine, pH 10.5 ± 0.1. The fluorescent unit (FSU) of the well contents was read on a CYTOFLUOR 4000™ fluorescent plate reader (PerSeptive Biosyslems, Framingham, MA) using 360 nm excitation and 460 nm emission filters. A four-parameter curve fit program was used to generate a standard curve, from which sample and control concentrations were interpolated. FSU readings were stable for at least 30 minutes at room temperature after 150µl glycine was added.

The ability to accurately measure VEGF in human plasma was assessed using several approaches. The effect of plasma on the assay sensitivity and performance was evaluated using rhVEGF165. Known amounts of rhVEGF165 were added to individual human plasma samples and the percent recovery determined as follows: (1) the amount of endogenous VEGF in the sample, determined from a parallel sample, was subtracted from the total amount of VEGF measured in the sample. (2) the 'recovered'VEGF value was then divided by the amount of VEGF added to the sample and multiplied by 100. The dilution linearity of rhVEGF165 added into individual human plasmas was also evaluated. In these studies, following rhVEGF165 addition, each plasma sample was diluted 1:10 in ELISA diluent followed by serial 1:2 dilutions in ELISA diluent. High and low matrix (standard) controls were prepared in neat human EDTA plasma (frozen). They were diluted 1/10 in ELISA diluent for a final concentration of 10% plasma.

Endogenous VEGF levels were measured in individual human plasma samples. Blood from normal healthy individuals was drawn into 15% K3 EDTA Vacutainer tubes (Becton Dickenson. San Jose, CA). The tubes were centrifuged at 2000×g for 20 min and the plasma was collected. Plasma samples were diluted 1:10 in ELISA diluent for use in the assay. The dilution linearity of endogenous VEGF in selected samples was also evaluated as described above.

The stability of neat human EDTA plasma was examined for three freeze-and thaw cycles. The plasma received a dry ice treatment followed by a gentle mixing in warm water in order to thaw. The data in Table 1 below demonstrate that there is no significant effect of quantitation of VEGF following freeze-and thaw treatments. Therefore, human EDTA plasma is stable for three freeze-and thaw cycles. hu EDTA Plasma Two-Site Multi-Site 32 22.11 78.84 Fresh 25.4 75.16 1 freeze-thaw 14.85 75.16 2 freeze-thaw 18.88 72.08 3 freeze-thaw mean 20.31 75.31 stddev 4.51 2.77 %CV 2.2 4 33 35.44 100.24 Fresh 35.87 101.71 1 freeze-thaw 35.44 101.71 2 freeze-thaw 39.19 101.71 3 freeze-thaw mean 36.49 101.34 stddev 1.81 0.73 %CV 5 1 34 17.51 78.18 Fresh 25.03 94.6 1 freeze-thaw 22.11 94.6 2 freeze-thaw 24.72 83.95 3 freeze-thaw mean 22.34 87.83 stddev 3.48 8.16 %CV 16 9 39 29.35 84.62 Fresh 31.11 81.39 1 freeze-thaw 28.02 75.16 2 freeze-thaw 31.11 71.51 3 freeze-thaw mean 29.90 78.17 stddev 1.50 5.93 %CV 5 8 38 23.43 83.95 Fresh 29.94 80.07 1 freeze-thaw 21.8 83.95 2 freeze-thaw 24.72 66.84 3 freeze-thaw mean 24.97 78.70 stddev 3.52 8.12 %CV 14 10 Mean 12 6 stddev = standard deviation CV = coefficients of variation

Approximately 20 replicates of the blank, 1. 2. 4 and 8 pg/ml standard, were assayed in the multi-site VEGF ELISA herein. The limit of detection was determined by the analyte (VEGF) concentration for which the measured mean FSU response minus two standard deviations was greater than the mean FSU response plus two standard deviations of the blank fluorescence emission (460 nm). Results in Table 2 show that the limit of detection is 8 pg/ml in ELISA diluent. Since plasma samples are typically diluted 1:10 to minimize matrix interference, as little as 80 pg/ml, or 1.6 pM VEGF can be measured in the original sample. replicates std 0 pg/ml std 1 pg/ml std 2 pg/ml std 4 pg/ml std 8 pg/ml FSUs FSUs FSUs FSUs FSUs 1 1103 1277 1351 1546 2216 2 1091 1382 1359 1617 2292 3 1103 1336 1413 1745 2241 4 1180 1328 1986 1770 2266 5 1235 1321 1382 1654 2216 6 1135 1382 1631 1735 2216 7 1180 1306 1328 1692 2266 8 1154 1125 1512 1654 2241 9 1079 1351 1366 1682 2279 10 1129 1559 1529 1678 2384 11 1129 1314 1445 1654 2266 12 1263 1413 1299 1617 2228 13 1079 1382 1336 1631 2216 14 1235 1284 1851 1716 2565 15 1135 --- 1711 1780 2266 16 1129 --- 1590 2077 2266 17 1017 --- 1445 1654 2279 18 1351 --- 1445 1740 2371 19 1079 --- 1546 1599 2318 20 2025 --- 1626 1663 2228 mean 1192 1340 1508 1695 2281 std. dev. 211 94 183 108 82 + 1 SD 1402 1434 1690 1803 2363 + 2 SD 1613 1528 1873 1911 2445 - 1 SD 981 1246 1325 1587 2199 - 2 SD 770 1152 1142 1479 2117

Two different preparations of rabbit polyclonal antibody against rhVEGF purified from the same rabbit but a different bleed were compared in the assay, using MAb 3.5F8 as the monoclonal antibody, and using 0.4 µg/ml of each type of antibody. The results, indicated in Fig. 1, show that both antibodies are suitable for use.

Rabbit polyclonal antibody elution was performed with glycine followed by guanidine and the resulting antibodies were used in the assay with the preferred conditions herein. Results in Figure 2 and Table 3 show that there is no significant difference between the two elution methods. However, the glycine elution seems to be slightly more sensitive. Comparison of normal human EDTA plasma samples as well as the High and Low Matrix controls show similar quantitation in both preparations. The guanidine peak is more tightly bound to the VEGF than the glycine peak.

Inter-assay and intra-assay precision was evaluated for the low and high matrix controls by ANOVA statistical analysis. Matrix controls were prepared by spiking rhVEGF into neat human EDTA plasma at low and high concentrations to fall within the assay range. Results show that the inter-assay variability (CV) ranges from 11-17% while the intra-assay variability ranges from 8-14%. The data is summarized in Table 4. Assays Name High (pg/ml) Low (pg/ml) Kn324p2 107.5 6.8 Kn324p2 111.2 11.6 Kn324p3 93.8 10.4 Kn324p3 96.7 11.2 Kn320p7 103.6 13.7 Kn320p7 103.6 13.7 Kn320p8 99.7 14.3 Kn320p8 102.0 14.7 Kn322p2 110.6 13.3 Kn322p2 102.0 --- Kn322p3 97.0 14.4 Kn322p3 103.9 13.9 Kn322p1 99.0 12.0 Kn322p1 95.2 12.3 Kn320p3 101.1 15.3 Kn320p3 98.0 14.5 Kn320p2 105.1 14.9 Kn320p2 100.5 14.2 Kn320p1 106.0 14.1 Kn320p1 103.4 13.9 Kn319p2 87.1 10.2 Kn319p2 100.2 9.3 Kn319p1 97.6 9.4 Kn319p1 99.6 9.1 Kn316p1 92.4 11.7 Kn316p1 97.9 11.3 Kn313p1 117.8 18.8 Kn313p1 111.4 16.2 Kn212p2 92.4 14.6 Kn212p2 92.4 15.7 Kn311p1 123.9 14.2 Kn212p1 103.8 13.9 Kn212p1 93.2 13.7 Kn331p1 106.2 12.9 Kn331p1 109.2 13.3 Kn331p2 99.8 11.9 Matrix Controls Intra-assay (% CV) Inter-assay (% CV) Low 14.0 11.0 High 8.0 17.0

Several samples in the past have shown non-linearity of increasing VEGF measured with increasing sample dilution. Therefore, the multi-site ELISA herein was tested for a hook effect (side-by-side comparison with the one- and two-site ELISAs), rhVEGF was diluted from 16 ng/ml to 1 pg/ml in assay buffer. Results (depicted in Figure 3) show that there is no significant drop in VEGF quantitation. However, a slight drop and plateauing effect can be seen from 512 pg/ml onward. Since the sample dilutions used in the multi-site ELISA assay herein give rise to concentrations less than 128 pg/ml, the hook effect is not a concern.

The procedures for determining coat maximization were the same as described above in the Methods section except that the concentration of either the PAb or the MAb 3.5F8 coat was varied. Specifically, for Fig. 4 the concentration of MAb 3.5F8 was varied from 0.4 to 4 µg/ml while keeping constant the concentration of polyclonal antibody (at 1 µg/ml). and for Fig. 5 and Table 5 the concentration of polyclonal antibody was varied from 0.4 to 4 µg/ml while keeping constant the concentration of MAb 3.5F8 (at 0.4 µg/ml).

While the 0.1 and 0.4 µg/ml concentrations of MAb 3.5F8 and of the PAb at the constant concentration of the other coat antibody were essentially the same in VEGF quantitation for the low and high controt, the upper limit of coat concentration (e.g., 0.4 µg/ml) is preferred to better the chances that the VEGF is captured. While the concentration of 1 µg/ml of MAb 3.5F8 and PAb increased the amount of VEGF measured in each case, such a concentration also gave higher background. Hence, the results show that the preferred concentrations for both capture reagents is about 0.4 µg/ml. MAb 3.5F8 0.4 µg/ml 0.4 µg/ml 0.4 µg/ml 0.4 µg/ml PAb 1 µg/ml 0.4 µg/ml 0.1 µg/ml 0 High matrix (pg/ml): 100.4 89.3 94.0 97.3 Low matrix (pg/ml): 22.4 9.7 10.1 5.4 Eight separate normal human plasma donors (pg/ml): 97.3 33.5 42.5 10.7 202.5 106.5 77.0 19.7 162.3 52.9 59.0 26.8 92.0 27.5 29.8 11.1 115.6 41.4 42.5 11.0 202.5 70.8 69.7 15.8 409.9 196.6 207.9 94.9 512.1 25.7 275.0 113.0

A pH profile was performed to determine whether changing pH of the assay buffer would increase or decrease recovery of VEGF in normal human EDTA plasma. Changing the pH could dissociate binding proteins or other complexes, if any, which would interfere with the MAb A4.6.1 detection.

The procedure for examining the pH profile of the assay was the same as described in the Methods section above except that for the sample incubation and biotin incubation, the assay buffer was adjusted using NaOH or HCl, resulting in assay buffers ranging from pH 4 to 9. A standard curve, a low and high matrix, and four normal human EDTA plasma samples were evaluated from dilutions performed using these varying-pH assay buffers.

Results in Fig. 6 and Table 6 show that there was no recovery of VEGF at pH 4 and 5. However, pH 6-9 revealed good VEGF plasma recovery with the assay control within an acceptable range. There was no significant difference in VEGF quantitation as a consequence of varying the pH of the assay buffer from 6 to 9. However, the preferred assay buffer is one with a pH of about 6.35 ± 0.1, which results in maximal VEGF binding and is appropriate for all dilution steps of the assay. pH 4 5 6 7 8 9 Normal Human EDTA Plasma pg/ml pg/ml pg/ml pg/ml 1 --- --- 198.9 122.0 173.3 204.0 2 --- --- 141.4 81.7 138.7 138.6 3 --- --- 240.7 150.3 220.9 243.5 4 --- --- 112.2 113.2 176.7 190.3 Controls pg/ml pg/ml pg/ml pg/ml High Matrix --- --- 104.5 105.0 93.8 124.0 Low Matrix --- --- 25.7 25.1 19.7 16.9

Approximately 85 pg/ml rhVEGF was spiked into neat human EDTA plasma and serially diluted to 1/10, 1/20, 1/40, and 1/80 and analyzed. The results, in Table 7 and Fig. 7, show that rhVEGF spiked in EDTA plasma showed linear correlation to expected concentration, with a coefficient correlation of 0.996. The percent difference between dilution-corrected concentration values determined for successive serial dilutions did not exceed a mean of 19% ± 7.5, as shown in Table 7. Normal Human EDTA Plasma (samples) [Measured] pg/ml Dilution Corrected Concentration % Difference S1 103 10 1026 --- 62 20 1237 21 35 40 1416 14 20 80 1599 13 S2 109 10 1088 --- 59 20 1176 8 35 40 1416 20 22 80 1788 26 S3 104 10 1039 --- 60 20 1202 16 32 40 1278 6 21 80 1677 31 S4 88 10 878 -- 52 20 1036 18 32 40 1278 23 19 80 1528 20 S5 93 10 926 --- 57 20 1136 23 32 40 1278 12 18 80 1433 12 S6 119 10 1192 --- 81 20 1612 35 47 40 1893 17 29 80 2298 21

Endogenous VEGF levels were measured in freshly-collected plasma from several normal healthy individuals. The individual human EDTA plasma samples were spiked with lowest, low, mid, and high concentrations of rhVEGF so as to fall within the assay range of the standard curve. Endogenous VEGF concentrations were determined and subtracted from the measured concentration to obtain comparison to the targeted spike. Results in Table 8 show that mean % recoveries were 99%, 113%, 106%, and 118% for the high, mid, low, and lower spikes, respectively.

An individual and two pooled male rat EDTA plasma samples were spiked with low, mid and high concentration of rhVEGF so as to fall within the assay range of the standard curve. Endogenous VEGF concentrations were determined and subtracted from the measured concentration in order to obtain comparison to the targeted spike (dilution control). Spikes were then diluted 1:2 in ELISA diluent to determine dilution linearity. Results in Table 9 show that mean percent recoveries range from 84-103% for the high, mid and low spikes that were greater than 6.25 pg/ml. [Expected] [Measured] [Measured] [Measured] Dilution Control pg/ml Male rat Pool 1 pg/ml % Recovery Male rat Pool 2 pg/ml % Recovery Individual Male rat pg/ml % Recovery Mean % recovery High spike 151 160 106 155 103 148 98 103 98 106 109 85 87 92 95 97 53 58 109 43 81 43 82 91 24 25 105 15 61 18 76 81 Mid spike 44 41 94 40 92 39 89 92 22 27 125 20 93 16 71 96 11 13 115 7 66 7 65 82 5 5 104 0 0 1 15 40 Low spike 20 18 88 19 93 12 62 81 10 14 135 9 86 6 63 95 6 6 102 1 18 0 0 40 3 2 152 0 0 0 --- 26 Lower spike 10 10 98 8 76 5 49 74 5 8 149 3 47 2 42 79 4 8 202 0 0 0 0 67 Endogeneous LTS 43 39

Rat EDTA plasma (2 male pools, 1 individual) was tested for linearity of dilution. Neat plasma samples were spiked with low (20 pg/ml), mid (44 pg/ml), and high (98 pg/ml) concentrations of rhVEGF and were serially diluted 1/10, 1/20, 1/40, 1/80 in ELISA diluent. Results in Table 10 and Figs. 8A-8C show that normal rat plasma samples dilute linearly following a minimum 1/20 dilution in the assay range of 8-128 pg/ml. Spike Rat pool 1 Rat pool 2 Rat 1 Mean R value High 0.996 0.996 0.999 0.997 Mid 1 0.999 0.994 0.998 Low 0.929 0.993 0.98 0.967

Eight Yorkshire pig EDTA plasma samples (four males and four females) were spiked with low, mid, and high concentrations of rhVEGF so as to fall within the assay range of the standard curve. Endogenous VEGF concentrations were determined and subtracted from the measured concentration in order to obtain comparison to the targeted spike (dilution control). Spikes were then diluted 1/10, 1/20. 1/40, 1/80 in ELISA diluent to determine dilution linearity. Results, shown in Figure 9A (females) and Figure 9B (males), and in Table 11, show that normal rat plasma samples dilute linearly following a minimum 1/20 dilution in the assay range of 8-128 pg/ml. Gender Coefficient of Correlation (R) Female 0.99718 Female 0.99917 Female 0.99998 Female 0.99958 Male 1 Male 0.99965 Male 0.99995 Male 0.99961 mean 0.99939 s.d. 0.00093491 %CV 0.094

This experiment was designed to determine if the multi-site ELISA herein could measure all the variants of VEGF. Figures 10A, 10B, and 10C show a comparison of the single-site, two-site, and multi-site ELISAs for VEGF, respectively. It can be seen by comparing these graphs that the multi-site assay herein is capable of capturing more VEGF variants.

The multi-site ELISA herein using PAb and MAb 3.5F8 as coat antibodies was compared to an ELISA using only MAb 3.5F8 or PAb as coat antibody for evaluating the amount of VEGF in normal human samples. The results, set forth in Table 12, show that the amount of VEGF detected in pg/ml was much higher for the multi-site assay than for the assay with PAb alone or MAb 3.5F8 alone. NHP Sample # Capture Reagent MAb 3.5F8 (Mean pg/ml) PAb to VEGF (Mean pg/ml) PAb to VEGF + MAb 3.5F8 (Mean pg/ml) 1 49 173 225 2 26 LTS 149 3 39 124 211 4 41 103 189 5 27 LTS 153 6 29 LTS 149 7 16 LTS 159 8 25 LTS 144 9 21 LTS 122 10 36 148 185 11 24 72 171 12 23 LTS 145 13 40 103 200 14 34 83 143 15 42 LTS 200 16 20 85 152 17 51 196 285 18 25 LTS 145 19 20 LTS 154 20 20 LTS 143 21 23 LTS 155 22 28 77 163 23 39 180 285 24 24 87 168 25 45 148 261 26 21 131 179 27 34 LTS 189 28 18 LTS 131 29 50 159 251 30 73 LTS 359 31 21 85 149 32 42 214 237 33 33 LTS 234 34 30 152 206 35 21 87 154 36 76 307 445 37 28 70 225 38 53 51 304 39 32 84 193 40 28 LTS 106 41 44 105 275 42 32 44 217 43 28 69 197 44 25 LTS 69 45 50 114 285 46 28 38 176 47 21 67 125 48 28 27 192 49 29 LTS 159 50 18 27 107 LTS = not detectable

Plasma and serum samples from normal human donors were analyzed by the two-site ELISA with MAb 3.5F8 as capture antibody and PAb A4.6.1 as detection antibody and by the multi-site assay herein using the PAb and MAbs and procedures noted in the Methods. The results, summarized in Figures 11A and 11B for plasma and serum respectively, indicate that the multi-site assay herein detects more VEGF in both types of samples than the two-site assay.

Samples from normal human donors and from donors with cardiac disease who were enrolled in clinical trials sponsored by Genentech, Inc. to evaluate efficacy of TNK, a t-PA variant, were analyzed by the single-site ELISA with MAb 3.5F8 as coat and detection agent, by the two-site ELISA with MAb 3.5F8 as capture antibody and MAb A4.6.1 as detection antibody, and by the multi-site assay herein using the PAb and MAbs and procedures noted in the Methods. Figure 12 shows the amounts of plasma VEGF in cardiac patients using all three assays, and Figure 13 and Table 13 summarize the amounts of VEGF in normal and cardiac patients using the two-site and multi-site assays by the mean amount of VEGF, standard deviation, % CV and s.e.m. The results indicate that the multi-site assay herein detects more VEGF in both types of samples than the two-site assay. Normal Donors Cardiac Patients Two-site Multi-site Two-site Multi-site mean pg/ml 32.61 192.40 37.54 279.23 s.d. 13.05 67.66 23.89 156.69 % CV 40.02 35.17 63.65 56.11 s.e.m. (Standard error mean) 1.84 9.56 5.34 35

Serum samples from non-small cell lung carcinoma patients were analyzed by the two-site ELISA with MAb 3.5F8 as capture antibody and MAb A4.6.1 as detection antibody and by the multi-site assay herein using the PAb and MAbs and procedures noted in the Methods. The results, shown in Figure 14, indicate that the multi-site assay herein detects more VEGF in lung cancer samples than the two-site assay.

Serum VEGF levels in normal humans and in patients with NIDDM (Type I diabetes) and IDDM (Type II diabetes) were measured using the two-site ELISA (MAb 3.5F8 as coat and MAb A4.6.1 as detection agent) described above. Figure 15 shows that the levels of scrum VEGF in NIDDM and IDDM patients were higher than in normal patients using this assay. Since the multi-site assay detects more VEGF than the two-site assay in other diseased patients, it would be expected that the multi-site assay herein would be suitable fordctccting elevated levels of VEGF in diabetic patients.

The two-site and muiti-site ELISAs were carried out as described above for normal human plasma samples. In addition, a multi-site was carried out using PAb to DNase rather than PAb to VEGF as coat reagent. All were at the 0.4 µg/ml concentration. Figure 16 and Table 14 show that the VEGF detected by multi-site VEGF assay is specific. Results from the ELISA using PAb to DNase plus MAb 3.5F8 show almost identical results as the ELISA using MAb 3.5F8 alone as capture reagent, with a slope of 1.04 (Fig. 16A). Human EDTA plasma MAb 3.5F8 (0.4 µg/ml) PAb to VEGF (0.4 µg/ml) and MAb 3.5F8 (0.4 µg/ml) PAb to DNase (0.4 µg/ml) and MAb 3.5F8 (0.4 µg/ml) High matrix (pg/ml): 147.9 147.4 119.2 Low matrix (pg/ml): 12.8 18.3 12.5 Seven separate normal human plasma donors (pg/ml): 61.1 164.0 59.5 33.8 120.4 48.5 34.6 174.3 64.5 47.4 104.5 102.1 28.3 70.4 79.6 67.6 113.0 71.8 61.1 105.9 65.4

TYPE OF ASSAY Immunoassay (Fluorimetric ELISA): mixture of a murine anti-VEGF monoclonal antibody (MAb 3.5F8) and a rabbit affinity-purified polyclonal antibody for capture and an anti-VEGF monoclonal antibody (MAb A4.6.1) for detection. STANDARD rhVEGF Reference Material VEGF or equivalent. Standard curve will be diluted in ELISA diluent: PBS/0.5%BSA/0.05%Polysorbate20/0.05% PROCLIN™ 300/5mM EDTA/0.35M NaCl, pH 6.35 ± 0.1 SPECIES QUALIFIED Human, Rat, Yorkshire Pig BIOLOGICAL MATRIX Serum, EDTA Plasma. ELISA diluent = Dilution buffer ASSAY RANGE 1-128 pg/ml in ELISA diluent QUANTITATIVE RANGE IN BIOLOGICAL MATRIX 80 pg/ml to 1280 pg/ml for human serum and EDTA plasma (1/10 minimum dilution)
240 pg/ml to 850 pg/ml for Yorkshire Pig EDTA Plasma (1/20 minimum dilution) Minimum Quantifiable Concentration Endogenous values as low as 120 pg/ml will be reported for information only with a %CV of approximately 60% in this range. Maximum Quantifiable Concentration 80 pg/ml to 1280 pg/ml for rat EDTA plasma INTRA-ASSAY PRECISION in Human EDTA Plasma Low 14% High 8% INTER-ASSAY PRECISION in Human EDTA Plasma Low 11% High 8% ACCURACY 6 normal human EDTA plasma samples were spiked with high, mid and low concentrations of rhVEGF. Mean % recoveries were: Low Mid High 106% 113% 99% 4 normal human serum samples were spiked with a high concentration of rhVEGF. Mean % recovery: High 113% 3 rat EDTA plasma samples were spiked with high, mid and low concentrations of rhVEGF. Mean % recoveries in the quantitative range were: Low Mid High 82% 91% 99% 4 female and 4 male Yorkshire pig EDTA Plasmas were spiked with high, mid and low concentrations of rhVEGF. rhVEGF quantitation below 12 pg/ml (endogenous VEGF measured) will be reported for information only (mean % recovery below 12 pg/ml is approximately 60%). Mean % recoveries in the quantitative range were: Low Mid High n/a 86% 107% SPECIFICITY rhVEGF. IGF, TNF NGF, hGH, IFN, IIbIIIa, rhuMAb VEGF, anti-VEGF MAb 3.5F8 were spiked into human EDTA plasma. Only the rhVEGF spike had measurable values above the endogenous VEGF serum levels. LINEARITY AND INTERFERENCE rhVEGF was spiked into 6 different normal human EDTA plasma samples. Samples were serially diluted ½ to cover the range of the standard curve. 4 dilutions were made from each sample, rhVEGF values obtained were plotted against I/dilution and the correlation coefficient (R²) of a linear regression analysis was calculated. Samples were linear across the range of the standard curve. n Mean R² SD 6 0.996 0.005 rhVEGF was spiked in 3 rat EDTA plasma samples at high, mid and low concentration. Samples were linear across the high and mid quantitative range. n Mean R² SD Spike 3 0.998 0.002 High 3 0.998 0.003 Mid 3 0.967 0.034 Low rhVEGF was spiked into 8 EDTA pig plasma samples and serially diluted ½ to cover the range of the standard curve. 4 dilutions were made from each sample. Samples were linear across the range of the standard curve. n Mean R² SD 8 1 0.001 SAMPLE STABILITY 5 normal EDTA plasma samples were tested for freeze-and-thaw stability. Human EDTA plasma is stable for 3 freeze-and-thaw cycles. Mean % recoveries: 1X 101% 2X 101% 3X 93% RUGGEDNESS/ ROBUSTNESS pH profile indicates that the multi-site assay can tolerate pH's ranging from 6-9. No quantitation of rhVEGF at pH 4 and 5. No effect on sample quantitation between a 1.5 and 16 hour sample incubation. (% difference in control recover: H ctrl 19%. L ctrl 16%). Assay performed at room temperature. No effect on temperature incubation (1 hour vs. overnight at 37°C) in VEGF quantitation of human plasma and serum samples. Mean % recoveries: 1 hour overnight serum 113% 121% plasma 115% 113% ASSAY ACCEPTABILITY Matrix Controls Westgard Multi-rules. ±2 SD from mean n (30%CV for Low Control) n/a Buffer Controls Standard Curve: Correlation Coefficient >0.994 Precision (%CV) <10% Accuracy (% differences) <20% Regression Parameters <20% CV Samples <20% difference between dilutions. SPECIAL COMMENTS Samples containing rhuMAb VEGF will interfere in the accurate quantitation of VEGF. Various samples were tested to improve sample linearity of endogenous VEGF (intending to release VEGF from possible binding proteins). Increasing amount of NaCl from 0.5-1.5 M in combination with varying pHs as well as the addition of glycine or KSCN (pretreatment) or heat inactivation did not improve dilution linearity or increase VEGF measured.

Little is known about the levels or the circulating forms of VEGF in normal individuals during growth, pregnancy, and old age or in pathophysiological disease states. Herein is described the development and characterization of a sensitive, high-throughput assay capable of measuring various isoforms of VEGF and their levels in human plasma. This assay represents an important tool for measuring VEGF levels in both normal individuals and in various disease states.

The multi-site VEGF ELISA herein can measure 165/165, 165/110, 121/121, and 110/110 VEGF variants equally well. With this assay, higher plasma VEGF was detected in normal donors (192 ± 68 pg/ml, n=50). For the same 18 cardiovascular patients, a significantly higher plasma VEGF above normal donors was detected (279 ± 157 pg/ml. n=18, p <0.001). See Fig. 13. Indeed, monoclonal antibody MAb 3.5F8 plus affinity-purified polyclonal antibody against an irrelevant protein did not generate any additional signal above that of MAb 3.5F8 alone. It is concluded that besides intact VEGF, other VEGF variants and isoforms are present in the circulation of both normal donors and cardiovascular patients. Ability to demonstrate that the receptor binding domain of VEGF is accessible for binding may be an important feature for any assay intended to understand the biological activity of VEGF in the circulation.

The fluorometric substrate, strep-β-gal/MUG, is preferred for use in the detection system so that the ELISA can detect endogenous VEGF levels in normal individuals. The use of this substrate and the determination of the best ELISA diluent resulted in much lower background absorbance, which was preferred to achieve the increase in the assay sensitivity.

The multi-site ELISA described herein is highly specific due to the choice of antibodies used for capture and detection. One of the coat antibodies, MAb 3.5F8. binds near the heparin binding region of VEGF (residues 111-165) and the other coat antibody, the rabbit polyclonal antibody binds VEGF. The detection antibody, MAb A4.6.1, binds in the KDR receptor binding region (residues 1-110) of the molecule, yielding a specific ELISA for VEGF.

The specificity of this multi-site ELISA will be important as the biology of VEGF is better understood. Keyt et al., supra (p.7788) have demonstrated that the different VEGF variants examined in this study have varying bioactivities in vitro. Knowledge of assay specificity will also be extremely important in evaluating clinical data and comparing data between laboratories.

Published repons (Kondo et al., supra (1994); Takano et al., supra (1996); Rodriguez et al., supra) have noted that serum VEGF levels were elevated in cancer patients. Considering that angiogenesis is a general phenomenon in solid tumor progression, and that expression of VEGF. a tumor angiogenesis factor, is observed in a wide variety of tumor cells of various origins, measurement of circulating VEGF levels has potential as a non-invasive diagnostic marker for a wide spectrum of solid tumors.

In conclusion, a sensitive ELISA that measures most molecular forms of VEGF has been developed. In accordance with the present invention, antibodies are raised in animals against human VEGF, with the C-terminal specific antibody being a monoclonal antibody and the whole-VEGF-specific antibody being a polyclonal antibody, preferably affinity purified. These two antibodies are used as coat antibodies (immobilized capture reagents) on a solid support such as microtiter plates. The antibody used for detection can be either polyclonal antibodies or monoclonal antibodies provided they are specific for the KDR and FLT1 binding domain regions of human VEGF.

Accurate and sensitive ELISAs like the one described herein are deemed important in helping to understand VEGF levels in various disease states. A better understanding of both VEGF levels and the dominant isoforms present in both normal individuals and in pathophysiological disease states will enhance knowledge of the role of VEGF in normal and pathologic angiogenesis.

While the invention has been described in connection with specific embodiments thereof, it will be understood that it is capable of further modifications and this application is intended to cover any variations, uses, or adaptations of the invention following, in general, the principles of the invention and including such depanures from the present disclosure as come within known or customary practice within the an to which the invention pcnains and as may be applied to the essential features hereinbefore set forth, and as follows in the scope of the appended claims.